Zhan-Wei Wu, Chih-Tai Tsai, Sao-Hsuan Lin, Yi-Syuan Ke, Jerry Chou 0001. MoE-Pipe: A Pipelined MoE Model Loading Framework for Reducing the Cold Start Delays in Serverless Inference. In Valeria Cardellini, Radu Prodan, editors, Proceedings of the 16th International Conference on Cloud Computing and Services Science, CLOSER 2026, Benidorm, Spain, May 19-21, 2026. pages 96-107, SCITEPRESS, 2026. [doi]
Abstract is missing.